4 October، 2022

Discussion of a doctoral thesis in the Faculty of Computer Science and Mathematics – Department of Mathematics, entitled:”  Numerical Optimization Methods with some Application in Improving  Artificial Neural Networks”

Discussion of a doctoral thesis in the Faculty of Computer Science and Mathematics – Department of Mathematics, entitled:” Numerical Optimization Methods with some Application in Improving Artificial Neural Networks”A PhD thesis for student Osama Muhammad Taher, under the supervision of Prof. Dr. Khalil Khader Abbo and Prof. Dr. Ibrahim Ahmed Saleh was discussed in the discussion room at the Faculty of Computer Science and Mathematics at the University of Mosul on Tuesday 4/10/2022The thesis dealt with deriving three new algorithms and then using these algorithms to improve artificial neural networks.The first algorithm is an improvement of the (HS) algorithm based on the convergence property of (Dai and Liao), and the convergence and regression behavior of the algorithm was studied and the numerical results were compared with the algorithms (FR, HS), the second algorithm is a new spectral conjugate gradient algorithm, and the convergence and regression behavior of the proposed algorithm was studied. The numerical results were compared with the (FR) algorithm, and the third algorithm is a conjugated gradient algorithm with a new parameter. The convergence and regression behavior of the proposed algorithm was studied and the numerical results were compared with the (FR) algorithm. As for the practical side, these new algorithms were used to teach FANs in three different types of applied problems (Simplefit, Bodyfat, Chemical) and compare the numerical results among them.Discussion Committee :1- Prof. Dr. Ban Ahmed Hassan / President 2- Prof. Dr.Huda Esam Ahmed / Member3- Prof. Dr.Shhbaa Ibrahim. Khalil / Member4- Associate Prof Dr. Ziad Mohammd A. / Member5- Associate Prof Dr. Salah Gazi Sharif / Member6- Prof. Dr. Khalil Khader Abbo / Advisor and Member7- Prof. Dr. Khalil Ibrahim Ahmed Saleh / Advisor and Member

Share

Share